Kohfeld et al. (1996) showed that oxygen isotopic ratios for N. pachyderma (s) are a useful proxy for water chemistry and temperature south of the Polar Front and suggested that the oxygen isotopic composition of encrusted forms reflects close to an average temperature of water above the pycnocline. At present, close to the drifts (figure F6 of Barker and Camerlenghi, 1999), the pycnocline is shallow and the average temperature above it is low (~-0.5°C). Late Pleistocene glacial average temperatures can have been little different. Grobe et al. (1990) presented data from N. pachyderma (s) from the eastern Weddell Sea margin that showed a correlation between carbon and oxygen isotopic variations through late Pleistocene glacial cycles, which the data reported here can examine.
The samples from Leg 178 are widely distributed and do not provide a detailed description of any single glacial cycle at either site (Fig. F1). However, they come from both laminated and bioturbated sediments, taken to represent both glacial and interglacial states, according to the assumed model of sedimentation on the drifts. This is borne out by the measured oxygen isotopic ratios. The high values, approaching 5.0 ppm, which are abundant at both sites (dominant at Site 1096) after 800-1000 ka, are difficult to reconcile with anything other than a low surface water temperature and large global ice volume. They come mainly from laminated sediments. Before 800-1000 ka, measured oxygen isotopic ratios are lower, reflecting perhaps smaller ice volumes and "interglacial" surface water conditions. Many more of these samples come from bioturbated sediments. In Figure F2, the oxygen isotopic values from both sites are plotted against the background of the detailed curve from benthic foraminifers at ODP Site 659 (Tiedemann et al., 1994). Precise correlation in timing is not intended by this comparison because of uncertainties in the estimated ages of the Leg 178 data (i.e., in the assumption of uniform sedimentation rates between magnetic reversals). However, it is clear from Figure F2 that the N. pachyderma (s) values reported here lie within the range of plausible isotopic variation (because low-latitude deep and bottom water is taken to have an origin in, and largely to retain the low temperatures and other properties of, high-latitude surface water) and reflect both glacial and interglacial conditions. Diagenesis could of course make such general coincidence merely fortuitous. Grobe et al. (1990) detect diagenesis (under reducing conditions) in samples from very shallow depths on the continental slope off Queen Maud Land. That diagenesis, however, leads to higher oxygen isotopic ratios. In Figure F2, the isotopic values decrease back in time, following the low-latitude curve, and low values occur at intervals throughout, so that tight limits are placed on the extent of any such diagenetic modification at Sites 1096 and 1101, and the general coincidence remains meaningful.
Figure F3 is a crossplot of carbon against oxygen isotopic ratios, which does not show the clear correlation between the two that was found by Grobe et al. (1990). Nevertheless, the range of carbon isotopic values in the Leg 178 data (leaving aside the three probably discrepant values, one from the lower part of Hole 1096B and two from the lower part of Hole 1101A, listed in Tables T1 and T2) is at least as great as that of the data reported by Grobe et al. (1990). It is clear that the existing data do not provide an opportunity for examining the level and significance of carbon isotopic variation south of the Polar Front through glacial cycles (e.g. Kohfeld et al., 2000), but a more detailed sample set from these sites would present this possibility.
Little can be said here about the abundance of N. pachyderma (s). Donner and Wefer (1994) in the Antarctic and Kohfeld et al. (1996) in the Arctic suggest that N. pachyderma (s) flux to the seabed is critically dependent on ice cover, even within polynyas. The relevance of the persistence of N. pachyderma (s) presence in the Leg 178 holes to a history of ice cover in this area during late Pleistocene glacial periods is uncertain, but that persistence seems likely to allow continuous sampling (particularly if only a small number of specimens would be required for each measurement).